swinv2-tiny-patch4-window8-256-dmae-humeda-DAV41

This model is a fine-tuned version of microsoft/swinv2-tiny-patch4-window8-256 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.9385
• Accuracy: 0.6818

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 3e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 256
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine_with_restarts
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 40
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	2	1.4874	0.4318
No log	2.0	4	1.3560	0.4432
No log	3.0	6	1.3117	0.4432
No log	4.0	8	1.2763	0.4432
No log	5.0	10	1.2602	0.5682
8.9996	6.0	12	1.2348	0.6023
8.9996	7.0	14	1.1982	0.5795
8.9996	8.0	16	1.1592	0.6136
8.9996	9.0	18	1.1142	0.625
8.9996	10.0	20	1.0682	0.6364
8.9996	11.0	22	1.0256	0.6477
7.429	12.0	24	0.9843	0.6705
7.429	13.0	26	0.9602	0.6705
7.429	14.0	28	0.9452	0.6591
7.429	15.0	30	0.9385	0.6818
7.429	16.0	32	0.9320	0.6705
7.429	17.0	34	0.9285	0.6477
6.2752	18.0	36	0.9239	0.6591
6.2752	19.0	38	0.9214	0.6818
6.2752	20.0	40	0.9206	0.6818

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0